Structure-based drug design strategies have been used to elucidate specific ligand recognition determinants and ultimately lead to the design of new small molecules for specific targets. The structure of the target receptor protein is the first requirement in structure-based drug design approaches. In the absence of a high-resolution structure of a given G protein-coupled receptor (GPCR), computational techniques like homology modeling can be used to build a 3D model. Briefly, the best structural template is chosen from the Protein Database (PDB) mainly considering sequence identity and similarity with the target receptor and the quality of the crystallographic structure (e.g. resolution). The sequence of the target receptor is aligned to the template structure using highly conserved residues and the known shared structural features to guide the automated or semi-automated alignment. The sequence alignment and the template structures are the input for the homology modeling. Energy minimization or molecular dynamics can be used to further refine and optimize the resulting 3D models. The ligand-receptor interactions can be identified by means of structure-based approaches, e.g. molecular docking. The information contained in the ligand-receptor complexes from the docking can clarify structural elements for molecular recognition and lead to a further optimization of the compounds and the design of new derivatives. The binding site of a given GPCR can be mapped for each class of small molecule ligands. Also, virtual searching of chemically diverse databases for novel chemotypes to bind to a given GPCR structure has been productive using both structure-based and ligand-based strategies. We have applied mutagenesis and homology modeling to the study of GPCR families for extracellular purines and pyrimidines and used the structural insights gained to assist in the design of novel ligands. These families consist of the adenosine receptors (ARs) and the P2Y (nucleotide) receptors. The structures of the human A2AAR were recently reported in the antagonist-bound state and in the agonist-bound state. These structures can reliably serve as modeling templates, with some adjustment, for other ARs due to the relatively high sequence identity between ARs (average 47% between human subtypes). We achieved the structure-function analysis of P2YRs by indirect means, using mutagenesis and homology modeling based on a template of the high-resolution structure of similar GPCRs, such as the newly determined structures of the P2Y1 and P2Y12 receptors. We have collaborated with one of the premier centers for X-ray crystallography of membrane-bound proteins, i.e. the lab of Prof. Ray Stevens (University of Southern California), to report the first X-ray structure of an agonist-bound A2AAR. We collaborate with Shanghai Institute Materia Medica to determine the structures of P2Y receptors, which display major differences in comparison to other GPCR structures. Automatic docking of known potent nucleosides to the agonist-bound A2AAR crystallographic structure, and to homology models of other subtypes, was performed, resulting in new predictions of stabilizing interactions and a structural basis for previous empirical structure activity relationships. We predicted binding of novel C2 terminal and 5' derivatives of adenosine and used the models to interpret effects on measured binding affinity and efficacy of newly-synthesized agonists. In order to investigate the validity of homology models and the inherent selectivity of a particular model in relation to close homologs, we constructed multiple homology models for the A1 adenosine receptor (A1AR), A2AAR, and the P2Y1 and P2Y14Rs, and docked, millions of commercially available compounds from chemical databases. The screens have yielded numerous potent and novel ligands. Structures of G protein-coupled receptors (GPCRs) have a proven utility in the discovery of new antagonists and inverse agonists, which may modulate signaling of this important family of clinical targets. However, applicability of active-state GPCR structures to virtual screening and rational optimization of agonists, remains to be assessed. We have studied adenosine derivatives and evaluated the performance of an agonist-bound A2AAR structure in retrieval of known agonists, and then employed the structure to screen for new fragments optimally fitting the corresponding subpocket. The binding models also explain a modest selectivity gain for some substituents toward the closely related A1AR subtype and the modified agonist efficacy of some of these ligands. The study suggested further applicability of in silico fragment screening for rational lead optimization for GPCRs in general. Small molecule screening libraries cover only a small fraction of the astronomical number of possible drug-like compounds, limiting the success of ligand discovery efforts. Computational screening of virtual libraries representing unexplored chemical space could potentially bridge this gap. Drug development for adenosine receptors (ARs) as targets for inflammation and cardiovascular diseases has been hampered by the paucity of agonist scaffolds. To identify novel AR agonists, a virtual library of synthetically tractable nucleosides with alternative bases was generated and structure-based virtual screening guided the selection of compounds for synthesis. We have introduced diazo groups that are subject to photoisomerization into AR ligands to study light's effects on the pharmacological activity. The putative A2A interactions of new hits identified using high throughput mass spectrometry were identified by modeling. The P2Y1 receptor (P2Y1R) is a G protein-coupled receptor naturally activated by extracellular ADP. Its stimulation is an essential requirement of ADP-induced platelet aggregation, thus making antagonists highly sought compounds for the development of antithrombotic agents. Here, through a virtual screening campaign based on a pharmacophoric representation of the common characteristics of known P2Y1R ligands and the putative shape and size of the receptor binding pocket, we have identified novel antagonist hits of M affinity derived from a N,N-bis-arylurea chemotype. Unlike the vast majority of known P2Y1R antagonists, these drug-like compounds do not have a nucleotidic scaffold or highly negative charged phosphate groups. Hence, our compounds may provide a direction for the development of receptor probes with altered physicochemical properties. The P2Y12R is an ADP-activated GPCR that is well validated clinically as an important target for antithrombotic drugs. Our new P2Y12R structure is the most predictive of the P2Y14R, a target for treating inflammation. Ligand docking to the P2Y12R-based model of the P2Y14R predicted poses of both reversibly-binding small molecules, consistent with observed pharmacology. We are using this approach to discover novel ligands of this receptor. We have modeled the interaction of a wide range of nucleotide and non-nucleotide ligands to the P2Y12 X-ray structures and a P2Y14R homology model to predict the way these important synthetic molecules are recognized in the binding site. These methods will aid in ligand discovery for this receptor, which has implications for treatment of thrombosis, pain, inflammation other diseases. We have also studied polypharmacology of rigid nucleosides containing a bicyclo3.1.0hexane ring system in place of ribose. Broad screening of hundreds of our compounds has identified off-target activity at other GPCRs (e.g. serotonin 5HT2B and kappa opioid receptors) and various channels. We have directed the SAR of newly synthesized analogues to optimize these newly found activities and to reduce the affinity at purinergic receptors. In this manner we have discovered new kappa opioid antagonists.